This module provides a solution for RuREBus: https://github.com/dialogue-evaluation/RuREBus
The model has won the second track is based on R-BERT (Wu et al. 2019): https://arxiv.org/abs/1905.08284
There are also some other models for RE and NER tasks but in main.py solution provided for R-BERT. However, main.py
can fall, because the module was copied from local repo and cleaned from experimental code (and there were some other
fixes of code style).
Nevertheless, if it's not working you can ask me, and I fix it. However, I think to improve this code in the nearest future to make it more usable.